Read me for: “Using Earnings Calls to Understand the Political Behavior of Major Polluters”

This replication directory contains files to replicate all analyses in the paper. 

The files are divided into two groups: data files and R code to reproduce results and figures.


NOTE: All code was run using RStudio Version 1.1.456 on R version 3.5.1 (2018-07-02) -- "Feather Spray”, 
Platform: x86_64-apple-darwin15.6.0 (64-bit)


# Data files (3):

* ecom-annual-data.csv: annual averages by firm for six indicators (climate, intlagree, natlaws, co2price, ccs, endff) along with average of the indicators (political).

* ecom-raw-data.csv: raw earnings call data including paragraphs from transcripts (para1, para3, para5) and coding for each indicator (Climate_science, Intl_agreements, National_laws, Carbon_pricing, Carbon_capture, End_fossil_fuel). Each row represents a paragraph-call-firm unit. Paragraphs are identified using CallIndex and ParaIndex along with dates (Day, Month, Year, Quarter). Word counts for each paragraph, call and firm are also given (ParaWC, CallWC, FirmWC).

* influencemap.xlsx: annual InfluenceMap grades for firms. See "notes" tab for more detail on variables.


# Code files (1):

* ecom-analysis.R: Replication code for generating Figure 1, Figure 2, Figure 3, and correlations reported in the paper.



